Errata

[This corrects the article DOI: 10.1200/JCO.2017.76.1825.].

National parks and other protected areas are integral to the conservation of biodiversity in tropical Asia (Squires 2014;Francis 2019). Bats are very important components of mammalian diversity in this region, including Vietnam, but many land management units lack intensive knowledge of the bat communities that inhabit them (Furey et al. 2010;Kingston 2010). Bidoup Nui Ba National Park (BNBNP) was established in 2004 as a high priority conservation area within Vietnam, is the fifth largest (66 000 ha) of the 33 national parks in the country, and spans a wide range of elevations from about 650 to 2300 m above sea level (Bidoup Nui Ba National Park 2019). One survey of small mammals at BNBNP conducted intermittently during 2002 to 2009 included preliminary efforts to document bats but noted that their inventory was "far from complete" (Abramov et al. 2009: 71). The objectives of our study were to more intensively determine the species diversity, reproductive phenology, and echolocation characteristics in the bat community of BNBNP. Studies of the bat communities of other national parks and nature reserves in Vietnam have been limited in number but have increased over the past 20 years (e.g., Hendrichsen et al. 2001;Furey et al. 2009Furey et al. , 2010Furey et al. , 2011Kruskop and Shchinov 2010;Minh et al. 2011;Thong 2015;Son et al. 2016;Tu et al. 2016).

Study area and forest categories
BNBNP is an isolated mountainous reserve located in Lam Dong Province on the Dalat Plateau, in the southernmost Central Highlands of south-central Vietnam (Fig. 1). Established in 2004, BNBNP has 90% forest cover composed primarily of montane evergreen (broadleaf) forest, with patches of coniferous forest and mixed broadleafconiferous forest (Tran 2011;Joshi et al. 2015). The area is influenced by a dry, cool season from November through March, and a warm, wet season from April through October (Pham-Thanh et al. 2019) with a mean annual rainfall of approximately 1870 mm, minimum temperatures of -0.1°C, and maximum of 31.5°C (Brodribb and Feild 2008;Tran 2011).
We tabulated descriptive forest categories at capture sites for each species of bat captured. We used five categories compiled from unpublished 2010 forest cover classes provided by the Vietnam Forest Inventory and Planning Institute, Hanoi with a scale of 1:50 000. We characterized forest within a 50 m circular radius around each bat capture site as agricultural and disturbed, broadleaf forest, coniferous forest, mixed broadleaf and coniferous forest (MBC), bamboo and mixed forest (BMF), or some combination of categories. Within tables we listed, in descending order of abundance, the forest categories present at all capture sites for each species of bat. We limited our forest cover results to descriptive totals within categories because most species of bats were seldom captured, forest cover was measured remotely and imprecisely, and vegetation analyses are lacking in most studies of southeast Asian bat communities for comparison. For analysis we provided a comparison of the proportions and 95% confidence intervals (CIs) of bats sampled that fell under the most frequent forest category (broadleaf forest) by families and most frequently captured species. We then examined for overlap of these CIs for proportions with the proportional availability of the most frequent category among all sites where netting and trapping occurred. We calculated proportions and CIs following the method of Newcombe (1998) using a correction for continuity.
We sampled bats at BNBNP near four field camps established during 2014-2016 in the dry (March) and early wet (May and June) seasons (Fig. 1). Camp 1 was located in the interior of the BNBNP on a foot trail 1.2 km from the nearest road at 12.09662°N 108.35684°E, elevation 1609 m, in broadleaf forest that was previously selectively logged in about 1980. Bats were captured at 23 of 32 geo-referenced sampling locations, all within 2.7 km of the camp and in the broadleaf forest remote sensing category (see below), at elevations ranging from 1541 m to 1786 m (X̅ = 1556 ± 359 m). We sampled at the Camp 1 area from 3 to 11 June in 2014, and from 21 to 26 March 2015. Camp 2 was in mixed agriculture-secondary forest habitat at 12.249424°N 108.436981°E, elevation 644 m on trails accessible from paved roads by motorbike. Bats were sampled at four locations in coniferous forest and in bamboo and mixed forest habitats at Camp 2, all within 0.3 km of the camp at elevations ranging from 650 m to 675 m (X̅ = 664 ± 11 m). We sampled at the Camp 2 area from 17 to 19 March 2015. Camp 3 was along a paved roadway in forested habitat near Giang Ly Forest Guard Station, 12.18241°N 108.67995°E, elevation 1454 m. We captured bats at 13 of 17 geo-referenced sampling locations at Camp 3 on 11 nights from 11-21 May 2015 and 16-23 March 2016; 14 locations were within 3.7 km of the camp, with one location 8.0 km north of the camp. Sampling elevations at Camp 3 ranged from 1094 m to 1656 m (X̅ = 1486 ± 136 m). Forest categories at netting sites around Camp 3 were primarily in broadleaf forest (11 sites), with other categories including agricultural and disturbed lands (two sites), mixed agricultural and broadleaf forest (two sites), broadleaf and conifer (one site), or bamboo and mixed forest (one site). A permanent plot of forest structure and tree biodiversity is located near Camp 3 (Hoa et al. 2018). Camp 4 was in mixed coniferous and broadleaf evergreen forest at 12.25294°N 108.63507°E, elevation 1057 m. The camp was accessible by foot trails into the park interior about 9.2 km (straight line distance) from the nearest road. Bats were captured on 14-17 March 2016 at 12 of 19 geo-referenced locations within 1.9 km of Camp 4 at elevations ranging between 1050 and 1108 m (X̅ = 1073 ± 20 m). Forest categories at sampling sites around Camp 4 were broadleaf forest (six sites), coniferous forest (three sites), combinations of coniferous with mixed broadleaf and coniferous forest (nine sites), or agricultural and disturbed land (one site).

Bat sampling
We captured bats using mist nets and harp traps set at ground level across trails, over small ponds and streams, or near edges of forest. Mist nets ranged from 3.0 to 18.0 m in length and were about 2.6 m in height, whereas harp traps ranged from 1.0 to 2.1 m 2 in area. Mist nets were set from two to 12 h nightly, whereas harp traps were left open all night. In June 2014 we sampled on eight nights for 458 m 2 harp trap h (m 2 hth; three traps per night) and 2724 m 2 mist net h (m 2 nh; 1-5 nets per night of three sizes ranging 6-12 m in length). In March 2015 we sampled on nine nights for 329 m 2 hth (0-3 traps per night) and 7903 m 2 nh (2-12 nets per night of six sizes ranging 3.0-18.0 m in length). In May 2015 we sampled on 11 nights, for 1233 m 2 hth (1-2 different traps per night) and 3750 m 2 nh (1-5 nets per night of four sizes ranging 6-13 m in length). In March 2016 we sampled for bats on 10 nights, deploying 995 m 2 hth (1-6 different traps per night) and 12 617 m 2 nh (6-14 nets per night of four sizes ranging 6-18 m in length). Total effort was 26 994 m 2 nh and 3015 m 2 hth, respectively.
We preserved voucher specimens in ~95% ethanol in the field for 12 hours and then reduced the ethanol concentration to 70%. We retained most bats as voucher specimens but released 53 on site (36 of these were Rhinolophus affinis). We categorized adult females as pregnant, lactating, post-lactating, or non-reproductive following standard field techniques for bats (Racey 2009). In cases where pregnancy was detected in voucher specimens we recorded numbers of visible embryos. Age was categorized as volant juvenile or adult based on fusion of the phalangeal epiphyses (Brunet-Rossinni and Wilkinson 2009). We verified species identifications on voucher specimens using external and cranial morphology. We relied on echolocation characteristics of three species of small Rhinolophus bats for confirmation of morphological identifications because identifications of vouchers from GenBank may be uncertain. Frequency of maximum energy (FMAXE) measurements of R. stheno at BNBNP were consistent with those from multiple areas in southeast Asia, but did not overlap with those recorded for R. microglobosus from the same regions (Francis 2008(Francis , 2019Hughes et al. 2010;Phauk et al. 2013). Rhinolophus pusillus and R. lepidus are difficult to distinguish morphologically and available DNA information is limited. Therefore we tentatively grouped morphologically similar bats with non-overlapping ranges of FMAXE averaging 93.7 ± 1.3 kHz as R. lepidus following several sources (e.g., Shi et al. 2009;Li et al. 2014;Raghuram et al. 2014), and bats averaging 100.1 ± 0.9 kHz as R. pusillus following others (e.g., Francis and Habersetzer 1998;Shi et al. 2009;Zhang et al. 2009).
We also took samples of liver or wing tissue in ethanol for DNA analysis for species verification. We verified identifications of eight species from nine samples using DNA analysis (Supplemental Table S1). A 685 bp fragment of the COI mitochondrial cytochrome c oxidase gene (DNA barcode) was amplified by using primers VF1d-VR1d (Ivanova et al. 2006). Tissue samples were extracted using DNeasy blood and tissue kit, Qiagen (Hilden, Germany). Extracted DNA from the fresh tissue was amplified by DreamTaq PCR mastermix, Thermo Fisher Scientific (Vilnius, Lithuania). The PCR volume consisted of 21 μl (10 μl of mastermix, 5 μl of water, 2 μl of each primer at 10 pmol/μl, and 2 μl of DNA or higher depending on the quantity of DNA in the final extraction solution). PCR condition was: 95°C for 5 minutes to activate the taq; with 40 cycles at 95°C for 30 s, 50°C for 45 s, 72°C for 60 s; and the final extension at 72°C for 6 minutes. PCR products were subjected to electrophoresis through a 1% agarose gel, 1st BASE (Selangor, Malaysia). Gels were stained for 10 minutes in 1× TBE buffer at 2 pg/ml of ethidium-bromide, and visualized under UV light. Successful amplifications were purified to eliminate PCR components using GeneJET™ PCR Purification Kit, Thermo Fisher Scientific (Vilnius, Lithuania). Purified PCR products were sent to Macrogen Inc. (Seoul, South Korea) for sequencing. Sequences generated in this study were aligned with one another using De Novo Assemble function in the program Geneious v.7.1.8 (Kearse et al. 2012). They were then compared with other sequences using the Basic Local Alignment Search Tool (BLAST) in GenBank.
All samples and voucher specimens were deposited in collections at the Department of Vertebrate Zoology, Institute for Ecological and Biological Resources at the Vietnam Academy of Sciences and Technology, Hanoi. Collecting methods, euthanasia, and specimen preparation followed guidelines for obtaining mammal specimens as approved by the Mammal Society of Japan (http://www.mammalogy.jp/en/guideline.pdf) and the American Society of Mammalogists (Sikes and the Animal Care and Use Committee of the American Society of Mammalogists 2016). Field work was carried out with the permission of the Ministry of Agriculture and Rural Development, Hanoi, Vietnam.

Analysis of species diversity
We measured species richness of bats at BNBNP as the total number of species captured (n). We calculated predicted species richness using the Solow and Polasky (1999) computation method in program Spade (Chao et al. 2016b) and a hypothetical increase in sampling effort that was double the total number of bats captured during 2014-2016. We estimated inventory completeness as the ratio of observed species richness to predicted species richness × 100%. We used program SpadeR (Chao et al. 2016a) to quantify species diversity as the maximum likelihood estimator for the inverse Simpson Index of Diversity (1/D) and expressed evenness of distribution of individuals among species as (1/D)/n (Magurran 1988).

Echolocation recording and analysis
We recorded echolocation calls of bats, including individuals that were prepared as voucher specimens, to compare results with the literature on echolocation of bats recorded elsewhere in Asia. Such comparisons may be useful aids to species identification, particularly for cryptic taxa that may be members of species complexes that are not yet well understood (Kruskop 2013;Francis 2019;Wilson and Mittermeier 2019), and can also indicate geographic variation in call structure within species (Ith et al. 2015(Ith et al. , 2016. Bats were recorded primarily as they were followed in flight in an enclosure made with mosquito netting (2 m high × 2.5 m wide × 6 m long), but also when hanging freely on the sides of the enclosure or while held in hand if flight did not occur. Recordings of bats in flight in enclosures or in hand are commonly employed for bat surveys in Asia (Kingston et al. 1999;Hughes et al. 2010Hughes et al. , 2011Kingsada et al. 2011), but measurements can be biased compared to recordings of free-flying bats. However, given low capture success for most species and the need for voucher specimens we did not release bats for recordings in the open.
We recorded and analyzed echolocation calls as WAV files using an Echometer EM3 digital ultrasonic recorder (Wildlife Acoustics 2016). The EM3 allows recording at sampling rates of 256 and 384 kHz (providing analysis of calls up to frequencies of about 192 kHz). We analyzed properties of recorded calls in Hanning windows using spectrograms, oscilloscope tracings, and power spectra features of Call Viewer software (Skowronski and Fenton 2008). We analyzed time and frequency characteristics for 12 calls per individual, selecting calls that provided the greatest amount of information. For bats with predominantly frequency modulated (FM), including FM/ quasi-constant frequency (FM/Q-CF) calls, we measured (all in kHz) start frequency, end frequency, frequency of maximum energy (FMAXE), midpoint frequency, bandwidth, and duration (ms). For bats with predominantly CF calls (including CF/FM and FM/CF/FM calls) we measured (in kHz) FMAXE, the frequency range of the preceding upsweep (FM rise) if present, and the frequency range of the terminal downsweep (FM tail), as well as the sound duration (ms). We did not measure interpulse intervals because of the confined recording context. We were interested in variation among calls within the species rather than variation among individual bats: for each measure, we provide mean ± 1 standard deviation (SD), 95% confidence limits (CI) for means, coefficients of variation (CV) as percent values (SD/Mean × 100), and ranges of calls. We compared consistency of our measurements with published values in the literature for the same species reported from elsewhere in Asia. Values in the literature are reported in a wide variety of ways, sometimes as single values or ranges with no other summary statis-tics, making such comparisons somewhat qualitative. We emphasized overlap of FMAXE measurements for each species reported elsewhere with those from BNBNP, but provide a detailed summary of previously published call measurements, references, and full summary statistics from our study as Supplementary Tables S2 and S3 to allow readers to make independent judgements. Dissimilar values may indicate possible geographic or habitat variation in echolocation frequencies within species, species identification issues, or evolutionary and taxonomic differentiation requiring future study.

Species richness, diversity, evenness, and general distribution
We captured 211 bats of 27 species in five families (Table 1). Thirteen species were considered rare: nine documented by single captures and four by two captures (Table 1). Eight species provided noteworthy distribution records (provincial or wider): Sphaerias blanfordi, Rhinolophus cf. marshalli, Hipposideros cf. swinhoei, Eptesicus pachyomus, Kerivoula dongduongana, Kerivoula titania, Nyctalus cf. plancyi, and Phoniscus jagorii; most of these also were rare and represented by just one or two bats captured (Table 1). Among the rare species, six (R. cf. marshalli, H. cf. swinhoei, E. pachyomus, K. titania, N. cf. plancyi, and P. jagorii) have seldom been reported in surveys anywhere within Vietnam (Table 1). Our molecular analyses confirmed identification of eight species with a high level of confidence (≥ 97.9% similarity with sequences published on GenBank) (Supplemental Table S1). One other sample (BNB030) was morphologically similar to R. lepidus (Supplemental Table S1), but additional confirmation is required because the R. lepidus and R. pusillus sequences within GenBank are assigned to specimens that may have uncertainties in their morphological identifications. Future DNA verification also would be useful to confirm our tentative morphological identifications of R. cf. marshalli, H. cf. swinhoei, E. pachyomus, and N. cf. plancyi.
Simpson's inverse index of diversity was 7.689 (CI 5.973,9.402) and evenness was 0.285 (CI 0.221, 0.348). A hypothetical doubling of numbers of bats captured would result in an additional 8.4 (CI 0.0, 18.9) added species, suggesting an inventory completeness of 75.8%. The index of evenness reflected the finding that relatively few species of bats were taken in abundance whereas many species were rarely captured. The six most abun-   Table 1). Two of the five species of pteropodids were taken only above 1000 m a.s.l. (Table 1). Most species of bats were captured at sites in forest cover categories that reflected the intensity of sampling within those categories. Remotely sensed forest categories at 74 sites (including sites where nets or traps were set but no bats were captured) at BNBNP were primarily in broadleaf forest (51; 68.9%, CI 57.0%, 78.9%), followed by coniferous with mixed broadleaf and coniferous (8), coniferous (5), agricultural and disturbed (3), agriculture and broadleaf forest (2), and one each in five other categories or combined categories. Except for pteropodids, the proportion of all bats captured were also mostly in broadleaf forest and within the 95% CI of the proportion of sites trapped or netted that fell into the broadleaf forest category (Table 2). This was true for bats in all families and for each of the eight species of bats captured most frequently (eight or more captures per species; Tables 1 and 2). The low proportion of pteropodids in this category corresponded to the greater proportions of captures of two species in the agricultural and disturbed cover category: Sphaerias blandfordi (6 of 8 or 75.0%, CI 35.6%, 95.6%) and Cynopterus sphinx (4 of 13 or 30.8%, CI 10.4%, 61.1%). The captures of most species of bats including pteropodids were sparsely distributed among a variety of categories and combined categories (Table 1).

Female reproduction: litter size and seasonality of birthing
We examined 94 females of 20 species for evidence of reproduction (Table 3). Pregnancies (n = 59) were observed in 14 species, all in March prior to the seasonal rains. Single embryos were recorded in all females examined except for twin embryos in three of four pregnant Pipistrellus coromandra and one Murina huttoni. Lactating females or volant juveniles were found in eight species during the rainy season (May, June), when no pregnant females were noted (Table 3). Capture of one volant juvenile pteropodid (Cynopterus sphinx) provided the only evidence for bat reproduction during the dry season.

Echolocation
We recorded calls of 19 species of bats. Twelve species were FM or FM/Q-CF emitting bats (Table 4, Fig. 2). Echolocation call characteristics for two of these species (Eptesicus pachyomus and Murina harpioloides) to our knowledge have not been previously reported. Calls of two other species (Kerivoula titania and M. huttoni) have been recorded previously only at single locations in Asia; calls of K. titania at BNBNP are consistent with those recorded in Thailand whereas those of M. huttoni are not consistent with calls recorded in China (Table 4, Supple-mentary Table S2). The calls of the remaining eight FM or FM/Q-CF emitting species are generally consistent with those recorded by others at most other locations in Asia (Table 4, Supplementary Table S2).
We recorded echolocation calls of seven species of CF, CF/FM, or FM/CF/FM emitting rhinolophid and hipposiderid bats captured at BNBNP (Table 5, Fig. 3). Variation in FMAXE was low compared to other acoustic measurements (CV 0.3-1.4%, Supplementary Table  S3). Comparisons with measurements of the same species made at many other locations in Asia are consistent with

Discussion
We captured 20 species of bats not previously reported from BNBNP. Abramov et al. (2009) documented 13 species of bats in a preliminary survey of the park. Seven of these species also were taken in our study, three of which were among the most abundant species in both studies: Rhinolophus affinis, R. lepidus, and Pipistrellus coromandra. Four other species (Arielulus circumdatus, Murina harpioloides, Myotis horsfieldii, and Scotomanes ornatus) were also taken in both surveys but were less common. No species of fruit bats were recorded by Abramov et al. (2009). Six species documented by Abramov et al. (2009) but not taken during our survey were Hipposideros armiger, Coelops frithii, Myotis muricola, M. phanluongi, Harpiocephalus harpia, and Miniopterus magnater. These species were rare in the earlier study, with five species documented by single captures and one captured twice (Abramov et al. 2009). Our sampling was biased towards species that use habitats lower to the ground, and species that forage at or above canopy level may have been missed.
At least 33 species of bats in five families have now been documented at BNBNP. This is a moderately high species richness among the parks and other nature reserves in Vietnam. Several studies have reported inventories of bats in parks and reserves that recorded 12 to 24 species (Hendrichsen et al. 2001;Son et al. 2016;Tu et al. 2016) although these varied in completeness and intensity. Studies of six other areas have documented 30 or more species of bats. Thong (2015) reported 47 species at the well-surveyed Cat Tien National Park in southern Vietnam, noting that seven of these needed modern taxonomic confirmation, whereas others may represent members of species complexes not yet recognized. Hendrichsen et al. (2001) reported 32 to 39 species at three national parks in karst areas of northern Vietnam: Phong Nha-Ke Bang, Cuc Phuong, and Pu Mat. Minh et al. (2011) documented 43 species in Pu Mat National Park. Records of 39 species were compiled for Hong Lien Son National Park and surrounding areas in northern Vietnam (Kruskop and Shchinov 2010). Furey et al. (2010) captured 36 species in five families with 694 captures at the karst-dominated landscape at Kim Hy Nature Reserve in northern Vietnam. The Kim Hy study was based on nearly 40 000 m 2 h combined net and harp trap sampling systematically throughout the year. Six more species were confirmed at Kim Hy Nature Reserve by additional studies, increasing the total to 42 (Furey et al. 2010). We have no doubt that further research will increase the number of known species from BNBNP. Predicted species richness estimates based on doubling of the sample size in our survey suggested an additional 8-9 species of bats [totaling 35 to 36 species without those documented by Abramov et al. (2009)], but with an upper 95% confidence limit of 19 additional species. Kruskop and Abramov (2011) suggested that four species of bats not documented by Abramov et al. (2009) were known from surrounding areas and were likely to occur in BNBNP. Two of these were detected in our survey [Cynopterus sphinx and Hipposideros gentilis, reported as H. pomona in Kruskop and Abramov (2011)] but two have not yet been captured within the park (H. larvatus and Miniopterus fuliginosus).
Similar to Cat Tien National Park but unlike reserves in northern Vietnam, BNBNP does not have an extensive karst substrate (Gillieson 2005). BNBNP therefore has fewer roosting opportunities for cave-obligate species of bats. However, the variation in roosting habits of Vietnamese bats reported mostly from caves is not well known. Some species of bats in Vietnam considered to be cave dwellers also will roost in rock crevices ( Rhinolophus affinis; Kruskop 2013), hollow trees, and foliage. Kruskop (2013) and Francis (2019) (Kruskop 2013). The faunal affinities of the species found at BNBNP are mixed, ranging from Indonesian and Malaysian elements to Himalayan and Palearctic species (Kruskop 2013). The higher elevation landscapes of BNBNP make this park suitable habitat for species that are tolerant of cooler temperatures; many of the species of bats at BNBNP are known to utilize forested habitats at mid to higher elevations elsewhere in Vietnam (Kruskop 2013). Future research should seek to determine if bat distributions follow elevational differences at BNBNP, or if our observations on elevational differences among species simply reflect biases in local abundance due to sampling effort or proximity of roosts at sampling locations. Similarly, additional sampling combined with extensive ground-truthing of vegetation where bats have been documented will be required to determine if some species of bats are found disproportionately in any particular vegetation type. Our measurement and analysis of forest cover was insufficient to conclusively make such distinctions.
Like many areas in southeast Asia, the Dalat Plateau has an annual pattern of predictable wet and dry seasons. In our area March marks the terminus of the winter dry season, with the summer rains beginning in April and persisting through October (Pham-Thanh et al. 2019). We found that the insectivorous bats at BNBNP show a reproductive phenology wherein pregnancy is seen in many of our samples from March, with lactation and the presence of volant juveniles but no pregnancies evident in June. This pattern of late dry season pregnancies and wet season production of young is common in insectivorous bats from many other tropical parts of the world (Fleming et al. 1972;Bernard and Cumming 1997;Racey and Entwistle 2000), with increasing evidence for similar patterns now accumulating from different study areas in Vietnam Kruskop 2013;Son et al. 2016). Future bat community surveys elsewhere in Vietnam should attempt to include both seasons in their sampling to further verify this pattern. Determination of litter size is also important demographic information rarely reported for many southeast Asian bats, as evidenced by our seemingly unusual finding of twin embryos in Murina huttoni.
Southeast Asian bats can contain complexes of morphologically similar species that may be distinguishable in part based on divergence in fundamental aspects of their ultrasonic calls (Francis 2008). For example, in Vietnam Rhinolophus pusillus and R. lepidus are thought to be part of "an extremely tangled" species complex that has not been fully resolved (Kruskop 2013: 124). At BNBNP morphological examination with limited DNA confirmation suggested that the frequency of maximum energy in echolocation calls do not overlap between these two species (Table 5, Supplementary Table S3). The patterns in echolocation traits we found, however, are not consistent within species across studies from different locations (Table 5), suggesting the need for further taxonomic and acoustic study of these two species. The possibility of cryptic species diversity based on echolocation calls also cannot be ruled out for some of the other species we recorded that show distinct differences across locations. Two possible examples from our study at BNBNP include the Hipposideros gentilis and R. cf. marshalli (Table 1). In other cases, significant variability in echolocation patterns can occur within species. For example, both geographic and habitat variability occur in the echolocation calls of R. affinis throughout its southern range (Ith et al. 2015(Ith et al. , 2016. The possibility of geographic and habitat variability in echolocation calls also cannot be ruled out for some of the species we recorded at BNBNP that were distinct from reports in the literature. Incorporation of descriptive patterns in echolocation calls will be a necessary adjunct to all bat community surveys in Vietnam and elsewhere in southeast Asia to help unravel the sources in variation of echolocation calls apparent from the literature.

Supplementary data
Supplementary data are available at Mammal Study online. Supplementary Table S1. Results of genetic comparisons using 685 bps of Cytochrome c oxidase subunit 1 (COI) Supplementary Table S2. Echolocation call characteristics of bats with frequency modulated (FM) and FM/ quasi-constant frequency (Q-CF) emitting calls recorded at Bidoup Nui Ba National Park, Vietnam, in comparison with calls of the same species as reported elsewhere in Asia Supplementary Table S3. Echolocation call characteristics of constant frequency (CF), frequency modulated (FM)/CF, and FM/CF/FM emitting bats recorded at Bidoup Nui Ba National Park, Vietnam, in comparison with calls of the same species as reported elsewhere in Asia